A constrained maximum-likelihood approach for efficient multipath mitigation in GNSS receivers

In this paper, we develop a new method for mitigating multipath effects in GNSS receivers, based on a constrained maximum-likelihood (CML) estimates of the multipath parameters. First, we apply a nonlinear transformation on the signal parameters to reduce the search space. Then, we define a new criterion for constraining the relative amplitude of the received secondary signal, and use the Lagrange multiplier method to solve the CML optimization problem. The resulting likelihood cost function has a unique minimum and yields to closed-form parameters estimates. The proposed method does not suffer from the correlation multi-peak problem, as for the standard discriminators, thus it can be used for any type of GNSS signal to mitigate both code and carrier phase multipath errors, including the new BOC signals. Numerical examples show that the CML approach gives a significant refinement to reach the optimal positioning solution.